|
|
|
|
Thursday, 16 April 2026 |
|
|
|
|
|
|
Blurring the lines of healthcare AI |
|
|
| If you asked me a year ago to break down the different areas of healthcare AI we’re covering, it might’ve looked like scribes, clinical decision support and a whole lot of administrative use cases, like handling prior authorizations. But now those segments are a lot more blurry than they used to be. |
| Yesterday, Abridge pushed further into AI medical search when it announced a partnership with JAMA and NEJM. That puts it squarely in competition with OpenEvidence, which also touts those academic relationships. OpenEvidence, for its part, is pushing deeper into scribe tools, and both Abridge and OpenEvidence have been moving into prior authorization work as well. |
| Now it’s less like scribes competing with other scribes, or medical AI search competing with other search tools. It's more like Abridge is competing with OpenEvidence, which is competing with general AI companies (OpenAI and Anthropic), which are competing with health IT infrastructure companies, which are competing with existing EHR companies... |
| We’ve been observing time and again that these healthcare-focused AI companies are vying to be the place clinicians go for all their AI tools. As it gets easier to build and deploy AI within healthcare organizations, companies will have to find other ways to stand out and be sticky. The lines are only going to get
blurrier. |
| This blurred-line competition is new, and it means that companies that were founded in one narrow lane are now merging over into the paths of their competitors. At the same time, we're seeing massive valuations across the space. The stakes are getting higher, and not everyone is going to come out a winner. |
| - Lydia |
|
|
|
|
|
|
|
Testing the models |
|
50 |
The percentage of chatbot responses that were problematic, when researchers asked popular chatbots everyday health and medical questions, according to a new BMJ Open study. Researchers posed questions to Gemini, DeepSeek, Meta AI, ChatGPT and Grok about cancer, vaccines, stem cells, nutrition and athletic performance. (Their prompts were designed to “strain” the models toward misinformation.) Problematic responses were ones that could lead users to ineffective or bad health outcomes.
|
|
|
|
|
|
This week in health Тech |
|
|
|
It’s been a while since we’ve heard any news about Cerebral’s ex-CEO Kyle Robertson. This week, the US asked a federal court for a preliminary injunction that includes imposing an asset freeze on Robertson’s latest company, Zealthy (h/t Sherwood News). In the case, first filed in April 2024, the US accused Zealthy of dangerous telehealth practices and fraud. |
|
|
Innovaccer is investing $250 million over three years to expand its platform, which uses coordinated AI agents to automate administrative healthcare workflows. The agents will work on a shared set of data and handle different steps, passing tasks along from patient access to care management and billing. |
|
|
|
|
|
|
|
|
|
|
Worldwide made. Thanks for reading.
|
|
|
|
|
|
|
|
